Expectations for small cap stocks have ebbed and flowed recently, especially since the presidential election. The Russell 2000 index is up about 16% since the end of October, but only 1.0% year-to-date through May, on a price-return basis. The number of earnings-per-share (EPS) estimate revisions has been increasing since the beginning of the year. However, the number of negative revisions has also been increasing—and there have been more of them.
The months ending February, March, and April of this year had the highest percentage of negative revisions in the last year. The period ending in May became more positive, yet remained below 50%. In fact, looking back further, the net revision ratio has been below 50% (more negative revisions than positive) in almost every period since mid-2011, as the chart below shows. (We are using rolling three-month periods, measured at the end of each month.) The grey bars indicate the net revision ratio, while the colored lines are the number of absolute positive (green line) and negative (red line) normalized EPS revisions.
Another noticeable trend is the declining number of total revisions (dotted black line). A possible explanation for the decreasing amount of total revisions could be the structural changes occurring in the industry, resulting in fewer analysts employed at sell-side research firms, but that is a topic for another discussion.
Focusing back to the negative ratio of positive-to-negative estimate revisions, one might conclude the outlook is decidedly bearish. Despite this backdrop, the Russell 2000 has performed fairly well over this period. Honing in on the period from mid-2011 where the revision ratio went negative for this extended period, the index has gained 71.9% (9.7% annualized) through May. (Chart below).
Furthermore, during this observation period, which consists of 67 rolling three-month periods where net revisions were mostly negative, the Russell 2000 had a positive return over the subsequent three months about 75% of the time with a median return of 3.4%. The next chart combines the net revision ratio (blue line) and the forward three-month return to the Russell 2000. We can see that as the revision ratio generally hovers below 50%, the orange bars are positive in 50 of the periods. This might suggest that investors do not care about the direction of change with respect to forecasting future earnings. A core philosophy of our traditional equity strategies is the belief that investors do react to this information and it is a foundational input into our quantitative models. So, why show this data? This view is at the aggregate level, whereas stocks we select for our portfolio focus on the portion of the universe where estimates are rising. Our research indicates this group of stocks (top 10% of analyst revisions) tend to outperform those with less favorable estimate revisions.
Some possible explanations for the Russell 2000’s positive performance despite this analyst behavior is that it has set the stage for positive earnings surprises (not coincidentally, this is another factor we utilize in our quantitative equity models). We have also seen that when stock market correlations are high, as they have been recently, there is less need to rely on the experts that issue forecasts, but when there is less certainty (more dispersion among stock returns) and the market is declining, there is a greater reliance on what analysts have to say about future earnings. Either way, this analysis is meant to be informative, and should act as a springboard for further discussions regarding analyst activities and behavioral anomalies. We, of course, will continue to delve into the behavior of analysts, as well as how we can leverage the observable reactions to these events.
Even though the net environment may be negative, there are opportunities to employ our behavioral models to the universe and identify relatively attractive stocks. Again, do investors care since the index is rising while estimate revisions are net negative? Internal research indicates the value of perfect foresight – knowing which companies were going to receive future positive revisions – is still powerful information. If you would have been able to predict these events precisely within the small cap universe, and were able to buy the top decile each month, it would have produced a 98.7% annualized return since 1996. This statistic is what keeps us believing in the process we follow in our pursuit of alpha by picking stocks that are expected to be beneficiaries of positive estimate revisions and earnings surprises. We have not seen any evidence that our quantitative models have experienced any degradation due to net negative analyst revisions over the last few years, and so remain optimistic with respect to our quantitative application to the small cap equity universe.